TPS is short for transactions per second.
It's a number, and also a common -- perhaps the most common -- performance benchmark for blockchains. But there's a lot of smoke and mirrors around how it's calculated (there's no one standard benchmark!), what it implies, and what the number actually means for dapp developers and end users!
TPS stands for transactions per second and is a common metric for blockchain performance. It’s often cited, and has become an important metric for blockchain comparison.
However, it is an extremely misleading metric as: (1) there is no universal method for obtaining TPS, and (2) it is trivial to boost the metric using simplified environments, and/or simplified transactions such as transfers.
A transaction is the generic term for an on-chain action that results in updated state.
Different types of transactions include a transfer, the execution or deployment of a smart contract, and data processing or updates. Transactions can range in complexity and time to execute, many of them requiring substantially more computing power and state resources than simple transfers.
A transfer refers to the transfer of digital asset balances from one wallet address to another. A transfer is also the simplest form of a transaction, which means all transfers are transactions. But not all transactions are transfers
TPS is misleading for two primary reasons: (1) there is no universal method for obtaining a TPS metric, and therefore (2) it is trivial to boost the metric using simplified environments, and/or simplified transaction data such as transfers.
For now, we’ll focus on (2). Recall that transfers are the simplest type of transaction. If one were to benchmark a blockchain’s performance based off of transfers alone, you could expect to get very high TPS numbers. Why? Because it’s executing transfers is trivial.
To cite a TPS metric borne of transfers alone is extremely misleading as it does not properly represent the full scope of transactions. Instead, this would yield a vanity metric, or a “theoretical TPS” metric. And notably, while the acronym technically works, TPS is not short for transfers per second.
Transfers are a subset of transactions, which means that all transfers are transactions. However, it’s important to note that not all transactions are transfers. And in fact, recent transaction history of live smart contract networks reveals that transfers are just a small subset of all transactions.
Theoretical TPS typically refers to the highest number of transactions a chain could possibly produce in the most ideal of scenarios—transfer-only transactions and very centralized nodes.
Theoretical TPS has been used as the basis of “max TPS” metrics for many years now. We’ve seen claims of 1,000 and 10,000 and upward of 100,000 max TPS metrics across various blockchains.
But these are theoretical for a reason! We don’t see these performance numbers in the wild. Production environments do not reflect ideal conditions which renders theoretical TPS a meaningless, misleading metric. A meaningful metric would be one that reflects “Real” TPS.
Real TPS is in direct opposition to the theoretical and hypotheticals, but is instead based on real-world, live-production data.
Using recent historical data gives a far more realistic, accurate view into true blockchain performance, and what users can expect to see!
Why? Because the types of transactions on-chain also change over time. The first several million blocks on Ethereum included very basic transactions, but more recent transactions are far more complex, calling intricate smart contracts with multi-step calculations like atomic transactions!
Monad gives you Real TPS. Monad uses real and recent Ethereum data, published on-chain to give actual, usable meaning to TPS.
10,000 transactions per second. 36,000,000 transactions per hour.
That's almost one billion transactions per day.
